auto-updater

Check installed community skill versions against source registries and present unified diffs for approval.

183|37|Updated May 15, 2026
One-click install
npx skills add https://github.com/zhou210712/claude-for-legal-ZH --skill auto-updater-zhou210712
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: auto-updater
Source: https://github.com/zhou210712/claude-for-legal-ZH/tree/main/legal-builder-hub/skills/auto-updater
Command: npx skills add https://github.com/zhou210712/claude-for-legal-ZH --skill auto-updater-zhou210712

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents unsafe or unwanted changes by ensuring community skill updates are reviewed via diffs and applied only after explicit approval.

Core Features & Use Cases

  • Update discovery by fixed commit SHA: Checks installed skills against the source registry using immutable commit SHAs (not tags/branches).
  • Diff-first trust review: Presents detailed unified diffs, with heightened focus on high-risk areas like hooks and MCP configuration.
  • GlassWorm QA gate for new versions: Re-runs skills-qa scanning before any update is applied, with hard rejection for certain regressions and enforced human approval for security-relevant diffs.
  • Freshness re-verification: Detects when a skill falls outside its freshness window and offers re-check options or re-verification when updates exist.

Quick Start

Run the auto-updater in notification mode to see available updates and the exact diffs for human approval.

Frequently Asked Questions about auto-updater

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I safely review skill updates before applying them to managed agents?

To safely review skill updates, you can generate unified diffs comparing installed community skills against source registries using fixed commit SHAs, ensuring changes are explicitly approved before application.

What is the best way to prevent unsafe changes to hooks and MCP configuration during skill updates?

Preventing unsafe hooks and MCP configuration changes requires enforcing explicit human approval for diffs affecting .mcp.json, hooks.json, and allowed-tools before any skill update is applied.

How does commit SHA comparison work for tracking community skill versions?

Commit SHA comparison tracks community skill versions by checking installed skills against the source registry using immutable commit SHAs instead of mutable tags or branches to guarantee version integrity.

Can I run automated QA scanning before updating installed skills?

You can run skills-qa scanning as a safety gate before applying updates, automatically rejecting regressions and enforcing human review for security-relevant diffs to ensure update safety.

How do I handle installed skills that fall outside their freshness window?

When installed skills fall outside their freshness window, the system detects the expiration and offers re-check options or re-verification to ensure the skills remain current and secure.

Does the auto-updater work with managed-agent workflows for controlled maintenance?

The auto-updater works with managed-agent workflows by providing controlled maintenance of installed skills, triggered by explicit commands like checking or auto-updater agent calls for safe updates.